CNN is a model which is currently widely used in acoustic scene classification. Sparse coding is a model which used to be very popular in acoustic classification field before deep learning technology is widely used. In this paper we combine these two models for acoustic scene classification. Specifically, the calibrated sparse representation based score is fused with the score obtained through CNN classification model for classification. Experimental results on TUT acoustic scenes 2017 dataset and LITIS Rouen dataset show that the proposed algorithm can make good use of the classification abilities of sparse coding and CNN.